EPUB 3.4 Is Here. What It Means for Reading Access

The comment period opened July 21, 2026. Three Candidate Recommendation Snapshots from W3C's Publishing Maintenance Working Group landed simultaneously — EPUB 3.4, EPUB Reading Systems 3.4, and EPUB Accessibility 1.2 (opens in new window) — and the digital publishing world now has until October 19th to respond with real implementations. That's not a long runway. But the fact that accessibility conformance requirements are bundled into this candidate recommendation cycle, rather than treated as an afterthought, signals something genuinely different about where this standard is heading.

For disabled readers — people who are blind, have low vision, have print disabilities like dyslexia, or use assistive technology to navigate digital content — this matters in ways that don't always make headlines. The EPUB format is how most commercial e-books, educational textbooks, and digital library content reaches people. When EPUB accessibility standards are weak or optional, the consequences aren't abstract. Screen readers encounter improperly structured content. Reflowable text breaks on magnification. Navigation landmarks disappear. The reading experience that sighted users take for granted becomes a series of dead ends.

What EPUB 3.4, Reading Systems 3.4, and Accessibility 1.2 Actually Do

The three specs work as a system, and understanding that relationship is key to grasping why this moment matters.

EPUB 3.4 defines how digital publications are authored and packaged — the structural rules that publishers follow when creating content. This is the authoring layer. EPUB Reading Systems 3.4 defines how the software that renders those publications must behave — what e-reader apps, library platforms, and browser-based readers are required to support. This is the rendering layer. EPUB Accessibility 1.2 sits across both, specifying the conformance requirements that verify whether a publication is actually accessible, and the metadata requirements that make accessibility features discoverable before a reader even downloads a file.

That discoverability piece deserves more attention than it typically gets. Under EPUB Accessibility 1.2 (opens in new window), publishers are required to include structured metadata describing what accessibility features a publication contains — whether it has text alternatives for images, whether it supports screen reader navigation, whether it has synchronized audio. This means a blind reader using a library catalog or a bookstore can theoretically filter for content that actually works with their tools before purchasing or borrowing. That's a meaningful shift from the current reality, where accessibility is often discovered only after the fact, through frustration.

The Gap Between Standards and Experience

Here's where cautious optimism has to stay cautious. Standards at the candidate recommendation stage represent technical consensus, not implementation reality. The compliance framework paradox that organizations face across accessibility standards doesn't disappear because a new spec is well-designed. Publishers, particularly smaller independent houses and educational content providers, face genuine capacity constraints in implementing these requirements. Reading system developers — the teams building Kindle, Kobo, Apple Books, and the open-source Thorium Reader — have their own implementation timelines that don't always align with W3C milestones.

The history of EPUB accessibility is instructive here. EPUB Accessibility 1.0 and 1.1 existed before this candidate recommendation. The standards weren't absent — implementation was. A 2023 survey by the DAISY Consortium, one of the primary advocacy organizations driving accessible publishing standards, found persistent gaps in how reading systems actually rendered accessibility features even when publishers had done the work to include them. The rendering layer failed the authoring layer. That's why EPUB Reading Systems 3.4 matters as much as the content spec itself — conformance requirements for the software that renders publications are where the rubber meets the road for actual users.

This also connects to a broader challenge in accessibility testing methodology. As our research on automated testing limitations documents, automated tools catch a fraction of real accessibility barriers. For EPUB content specifically, the gap between what validation tools flag and what a screen reader user actually experiences can be substantial. The EPUB Accessibility 1.2 specification's conformance framework will only be as strong as the human evaluation practices that accompany it.

Who Is Affected — and Who Needs to Act

The populations with the most at stake in this standards cycle are specific:

• Students with print disabilities who rely on accessible textbooks — a population that has fought for years through legislation like the Marrakesh Treaty (opens in new window) and domestic accessible instructional materials mandates
• Blind and low-vision readers who use screen readers and refreshable braille displays to access commercial fiction and nonfiction
• Public library patrons who access digital collections through platforms like OverDrive and Libby, where accessibility metadata could transform how they discover usable content
• Older adults with age-related vision changes who depend on reflowable text and font scaling

The Southeast ADA Center (opens in new window) and peer regional centers have documented how digital content barriers compound existing access gaps, particularly for disabled people in rural areas where physical library access is already limited and digital collections are the primary option.

For practitioners — publishers, platform developers, accessibility auditors, and procurement officers at libraries and educational institutions — the October 19th comment deadline is a real action point, not a passive milestone. The W3C's GitHub issues process for these specifications is open and public. Organizations that implement these specs before they reach final Recommendation status, and document what works and what doesn't, shape the final standard. That's the mechanism the candidate recommendation process is designed to use.

Practical Standards Reference

| Specification | Layer | Key Obligation | Primary Citation | Practitioner Takeaway | |---|---|---|---|---| | EPUB 3.4 | Authoring | Structured, semantically enhanced content packaging | W3C EPUB 3.4 CR (opens in new window) | Publishers must implement semantic HTML structure, proper heading hierarchies, and image alt text at the authoring stage | | EPUB Reading Systems 3.4 | Rendering | Conformance requirements for user agents rendering EPUB content | W3C EPUB RS 3.4 CR (opens in new window) | E-reader developers must support accessibility features authors implement — rendering failures are spec violations | | EPUB Accessibility 1.2 | Conformance + Discoverability | Accessibility verification and metadata requirements | W3C EPUB A11y 1.2 CR (opens in new window) | Publishers must include structured accessibility metadata; procurement teams should require it in vendor contracts | | WCAG 2.1 AA | Web content baseline | Success criteria for perceivable, operable, understandable, robust content | W3C WCAG 2.1 (opens in new window) | EPUB Accessibility 1.2 requires WCAG 2.1 AA conformance as its content baseline — these standards are linked |

What the Comment Period Requires From Practitioners

The candidate recommendation period is where implementation evidence gets collected. W3C needs real-world deployments to confirm that these specifications are technically sound and implementable before they advance to full Recommendation status. Organizations that engage now — submitting implementation reports, filing GitHub issues on the specifications, and testing against the conformance criteria — have direct influence over the final standard.

For libraries and educational institutions operating under Title II of the ADA (opens in new window) and Section 508, the trajectory here is toward greater specificity in what accessible digital content actually requires. The standards fragmentation problem that has complicated compliance for years is not solved by EPUB 3.4 alone — but a well-implemented EPUB Accessibility 1.2 gives procurement officers a concrete, citable standard to include in vendor requirements.

The reading experience that disabled people deserve has a clearer technical path forward than it did before July 21st. Whether that path gets walked depends on who shows up to implement these specifications — and whether the community most affected has a voice in how they're finalized. Comments are open through October 19th at the W3C GitHub repositories for each specification. That's the most concrete next step available right now.